Examination of Witnesses (Questions 421
- 439)
WEDNESDAY 16 JULY 2008
Professor Sir John Bell and Professor Sir Alex Markham
Q421 Chairman:
Could I welcome you both first of all, Sir John and Sir Alex,
and thank you very much for making time to come and see us. We
regard this session as really very important; please feel free
apart from the questions to comment on any other issues that you
may wish. Can I also welcome the public at the back; you will
find information relating to the interests of the Members, please
help yourself. We are webcast so we have to keep private conversations
to the minimum or as quiet as we can, but the best is to avoid
it. I will invite you to make any initial comments that you have,
either of you, but the first time you speak it would be helpful
for the record if you would introduce yourselves and, if you do
wish to mention whom you represent, that is fine.
Professor Sir John Bell: Thank you very much,
My Lord Chairman. I am Sir John Bell and I guess I hold three
jobs that are relevant to this: one is the Regius Professor of
Medicine in Oxford, with a background in genetics, particularly
in immunological disease, the second is the President of the Academy
of Medical Sciences and the third is the Chairman of OSCHR. Alex,
why do you not introduce yourself?
Professor Sir Alex Markham: I am Alex Markham,
I am the Chairman of the OSCHR Translational Medicine Board, I
am also the senior responsible owner of a programme in the NHS
Connecting for Health activity called the Research Capability
Programme, and my day job is that I am the Professor of Medicine
in the University of Leeds.
Q422 Chairman:
Thank you very much. Does either of you have any opening comments
you want to make?
Professor Sir John Bell: I just wanted to start
by saying that this particular inquiry is extremely timely because
we are at one of those inflection points that you get in medicine
every so often where there are very significant opportunities
to apply this methodology in a patient setting and having a clear
strategic view about how one would go about that at this moment
in time I think is very timely, so I must say I welcome the inquiry.
Q423 Chairman:
Sir Alex, do you have anything to add?
Professor Sir Alex Markham: I would only echo
what Sir John has to say but I would also perhaps add a word that
modesty prevents him saying himself which is that this inquiry
takes place against a backdrop of one or two years of successful
work in changing the landscape in the UK in the way that we do
manage significant change in medical research. OSCHR has had a
very successful first year or so of activity. Some of the issues
that you have already had brought to your attention about the
challenges in genomic medicine, particularly things like informatics
and the way that new discoveries are pulled through in the so-called
translational setting to be applied for patient benefit. There
is an awful lot of work being done in those areas and it would
be good to reassure your Committee that that activity is already
taking place. You can have a good look-see at what has been achieved
so far that might be of relevance to your deliberations.
Q424 Chairman:
Thank you very much. That actually leads into two questions that
I have. One is probably more directed to you, Sir John, and the
other one is probably more directed to you, Sir Alex, but feel
free to comment, both of you, on both questions. The first one
to you, Sir John, is we have had a lot of evidence now about the
way the science of genetics and genomics is going, particularly
in relation to common diseases, and although the progress of science
is extraordinarily rapid there is some concern as to how much
of this is going to translate into healthcare, particularly the
understanding of common diseases. What is the evidence that these
advances are important and where do you expect these advances
to lead us and in what timescale, particularly the use of these
tests in understanding and identifying common diseases.
Professor Sir John Bell: I suppose I should
say that I am on record in a paper I published in the British
Medical Journal in the mid Nineties saying that the timeframe
would be five years. That timeframe lapsed about three years ago,
so I am on record of getting the timeframe wrong once before—that
is a health warning. There have been really dramatic changes in
the pace at which the discovery process in this arena has developed,
particularly in the last couple of years. We cannot really do
anything unless the discovery activity is delivered. You will
have heard the information about large whole genome association
studies, the abundance of new genetic variants but also the discovery
of new technologies that make this much easier to apply in a more
routine setting that generate more specific and more substantial
data that could be available to us. In my view, therefore, there
is likely to be an incremental growth of its use in common complex
traits, but it will not all happen overnight. You have to differentiate
the use of these tools for prediction of asymptomatic people early
in life, perhaps at birth, which I think may happen but is not
going to happen in the short term. There are a whole set of things
you have to put into place for that to work, and separate those
from some of the applications of genomic technology which can
be applied in an active clinical setting, which would involve
stratification of patient populations, identifying responders
and non-responders to particular types of therapy, monitoring
therapies and genetic tools. All those things are likely to happen
in a very short timeframe, particularly as the incentive to do
it is enormous and I think there will be a huge amount of effort
to try and make that deliver in a relatively short timeframe.
If you want me to put a number on it, certainly within five years
there is going to be a lot of activity in this arena using genomic
technologies, broadly defined, for common disease.
Q425 Chairman:
Are we investing enough in the big science related to genomics
to make this happen?
Professor Sir John Bell: How much is enough
and of course scientists could always claim they could always
use twice the budget, but the truth is the UK is really, really
competitive in international terms. In fact, you might argue that
we are now entering an era where we are actually right out in
front so if you look at the output of large case control studies
there is no doubt that the best data and the most data came out
of the UK studies. If you look at studies like the UK Biobank
project, which is essentially a large genetic project in common
disease, everybody else in the world would die to have a project
as good as that. So when it comes to applying genetic technologies
in a setting where you have well-characterised patient populations,
and people who are willing to participate in research in very
large numbers, the UK is uniquely positioned. Of course I would
say we need more money, my job is to say we need more money, but
the truth is we are doing really well in the basic sciences.
Q426 Lord Taverne:
You are optimistic about this. We have had some more pessimistic
views about the short term benefits we are likely to see, really
based on two propositions, that either the extra risk in the case
of common diseases that you discover is relatively small—it
might be a 30 per cent increase but from a very small base—or
else in cases where there was a cumulative build-up of risks the
actual cohorts affected were very small so that the prospect of
early benefits were really not all that great.
Professor Sir John Bell: Let me just go back
to a comment I made earlier, and that is I think if you focus
your attention on the snips associated that have emerged from
whole genome association studies and you say that is genomic medicine,
then you would be very badly wrong. That may be genomic medicine
but there is a whole set of other tools that are having an impact
today. You will have heard about the comparative genome hybridisation
arrays—CGH arrays—which are now systematically being
applied to a whole range of disorders in clinical genetics laboratories,
including all the major leukaemias, many of the common cancers
and a number of developmental disorders. It is not being applied
to diabetes but it is being applied to a whole range of other
diseases and those things are starting to happen in clinical practice,
and it will completely replace the whole area of cytogentics which
has historically been a big activity in the NHS. That has gone
because no one would sit at a microscope and look for translocations,
deletions, and duplications when you have the other tools that
allow you to do that much more systematically. So there are really
good examples of how those things are already impacting. The whole
issue of stratification might relate to the snips that are emerging
from whole genome association but it might not, you may not need
those because a lot of it can be generated from direct sequence
data, from transcript profiling data, from all the other tools
that are available to this field. I think, therefore, that the
concern about predictive testing and the relevance of the whole
genome association data from predictive testing is fair. We do
not yet know what that will do if you add up all those very small
risks, what it will do in a population to identify people and
what you would do if you had that information anyway. That is
why I said that predictive testing in asymptomatic people is actually,
in my view, the last thing that is likely to play but all the
other things are beginning to play now and are likely to continue
to play actively, so I guess it really depends on where on the
landscape you want to place your bets.
Q427 Chairman:
Have you got enough facilities for C-fast sequencing?
Professor Sir John Bell: As you will have been
told, the revolution of the last two or three years has been the
new generation to sequencing and it is truly revolutionary because
the capability of the new machines is allowing geneticists to
generate data—interesting sequence data—at a rate
and in different dimensions that we never dreamed we would be
able to do. At the moment there are three common platforms that
people are using and some are better for clinical diagnostic use
than others but, for example, in Oxford we have just installed
two 454 machines with the intent of using them in a purely translational
mode to actually facilitate clinical diagnostics of all kinds.
One of the crucial issues is that you can talk about all kinds
of surrogates of information which is genetic information that
you might like to have. This would include snip variation, transfer
profiling variations, and all kinds of intermediate measures from
which you infer a particular outcome. The truth is the ultimate
sort of data that you can generate is sequence data. If you can
get sequence data, the chances are you may not need any of the
rest of it, so the ability to generate large amounts of sequence
data becomes very central. Let me also warn you that this is the
first of several waves of new sequencing technology and I am familiar
with at least one other platform which is a UK-based platform
which will be capable within five years of generating, for around
$1,000, a whole genome of an individual in a very short timeframe.
That is the next wave of technology. One has to be conscious of
the fact that we are going to be at the receiving end of almost
unlimited amounts of genetic information and the real questions
are what useful information can be gleaned from the basic data
and is it useful in a clinical setting, does it have clinical
utility. That is a big hurdle and one which we have not been very
well-positioned to address up to now, I have to say. Secondly,
what is the best structure to generate that data? Is it better
just to sequence everybody at the beginning and have them carry
it around on a chip, or is it better to do it disease by disease?
Is it better to do it in regional labs? Is it better to do it
in local labs? There is a whole cascade of issues that emerge
once you have the evidence of clinical utility.
Q428 Chairman:
Alex, do you have any comments?
Professor Sir Alex Markham: Maybe it would be
helpful if I went to other end of the spectrum here and in anticipation
of a cascade of development, both technological and scientific,
do we have any structures in place that are thinking about how
we might handle this? In the cancer world we have been working
along these lines since about 2004. There is a thing called the
National Cancer Research Institute Informatics Initiative with
all of the funders, including the industrial sector, looking at
how we will pull together this massive cascade of data as it emerges,
anticipating that it will probably be several orders of magnitude
bigger than we envisage today, and also issues like how we integrate
it with the USA and other parts of the world. There is some very
solid thinking that has gone into that and has all the major research
funders engaged in it which gives me a level of confidence that
we can deal with the sorts of challenges that John is throwing
up for us. The other piece of work again in the cancer arena is
that for the first time we are now looking at the data that is
collected in cancer registries; that has been a statutory requirement
since 1960 and we have wonderful data on everyone who has had
cancer in Britain since that time—nobody has ever looked
at the data. This work was the source of the news items of two
or three weeks ago, highlighting the difference in death rates
in the north of England from lung cancer versus the south of England,
and it is really a bit scandalous that that sort of thinking and
information was not looked for from the public eye. The cancer
world is therefore doing a lot of thinking about this and always
the UK Clinical Research Collaboration activity builds on that
kind of pioneering work in the cancer sector. All of the efforts
that are now going in through OSCHR, through its e-Health Board,
and through the Department of Health through the Connecting for
Health process with this new programme called the Research Capability
Programme, are all about setting up the systems that will enable
us to handle this tidal wave of information under the right limitations
of good governance. It is to ensure all of those vital components
of this game, and that patient confidentiality is maintained,
that this is not on the front page of the newspapers on a daily
basis when data is lost. That challenge therefore is equally as
important as the technical side.
Q429 Chairman:
In your role with Connecting for Health do you think the NHS has
got the capacity to do that, particularly about confidentiality
of patient information?
Professor Sir Alex Markham: If we do not then
I guess it will be my fault. The problem with the confidentiality
landscape, as you will know from many reports, both from this
week with the Walport and Thomas report on data sharing, back
through reports of the Academy of Medical Sciences, the reports
pile high but the solutions are a bit more elusive. We have so
many bodies that see themselves as the ultimate guardian of the
patient's best interests that it is actually quite difficult to
pick your way through the landscape. You can name the Information
Commissioner, the National Information Governance Board, the Patient
Information Advisory Group, the General Medical Council, the British
Medical Association—all of these bodies consider that they
have something fundamental to say about that issue. It might not
have escaped your Lordships' notice that actually none of those
organisations actually can claim to be the last port of call.
One of the biggest workloads for me is at the moment working very
hard with all of those bodies to very simply say "I am not
here to do something cavalier with the patient information",
but there is nothing more unethical than preventing ethical medical
research taking place, and some of your bodies ought to give that
just a little more weight in their deliberations. There is a real
challenge there and I do not under-estimate it for a moment.
Q430 Baroness O'Neill of Bengarve:
I want to put in a very quick supplementary which is just this,
who is the last port of call on the model we have put ourselves
in on the protection of personal data, could it be Parliament
and the particular character of the Data Protection Act 1998?
Professor Sir Alex Markham: I think it is that
and, obviously, the owner of patient records is actually I believe—Lord
Warner will correct me if I am wrong—the Secretary of State
for Health. There is a mechanism, therefore, but I am not sure
if this personal view is correct. A lot of structures have come
into existence as a result of scandals like Alder Hey, like Bristol,
like the Shipman fiasco and they do not fit together coherently.
They were all done with best endeavours, with absolutely the right
desire to preserve patient confidentiality and hold the medical
profession to account, but the result is a very fragmented landscape
that somehow we need to pick our way through and I do not pretend
to have a complete solution to that. We are going down a route
of primarily using anonymised data and of building robust systems
to use pseudonomised data, i.e. to remove all identifiers from
public clinical records. There are going to be some cases, particularly
in the genetics arena, where it will be important to be able to
go back to a patient who is found to have a previously unanticipated
problem. That whole question of how do you get consent for consent
to approach somebody who does not know they have a problem at
the moment is very difficult.
Q431 Lord Warner:
Sir Alex, could I just try and get something clear because it
is a very complicated but very significant area. Of what I would
call two of the "pillars" on which we have operated
so far, pillar one has been that the use of any material which
has human cells in it is actually something on which you need
to consult and require, in most cases, the approval of the patient
from which those cells are extracted. That is underpinned by the
legal interpretations that have been made of the European Convention
on Human Rights as enshrined in the Human Rights Act in this country;
that has been the first pillar. The second pillar has been that
for research purposes you need consent but you reassure people
by giving some assurances about anonymity in the way it is used.
Are you saying that those pillars which have underpinned most
of the way we have acted over recent decades—they have been
more refined but they have been the basic principles—are
a problem or are you saying that the mechanisms which have been
used to interpret those principles have got into a mess and it
is the mechanism that you need to get right, not the principles
themselves?
Professor Sir Alex Markham: Absolutely the latter,
the principles are fine but the mechanisms have got bent out of
shape. Maybe that is a personal view but I think a lot of people
will share that. I think as a consequence of the level of confusion
instead of taking, based on those two fundamental pillars, a clear
message to the public which says you may be interested to participate
in this—I will give you an example if I may, just jumping
aside slightly. The experience with UK Biobank was that there
was an enormous debate as to whether it was ethical to write to
individuals, uninvited, and ask them whether they would consider
participating. The question was could you go to general practice
records, which are actually very accessible electronically? It
is not something the public really realises, just how accessible
GP records are—in the most ethical ways. Could you write
to people and say there is a programme going on and we are trying
to recruit 500,000 people to do this thing, and most people we
think would think that is a good thing to do. There was a huge
debate as to whether you could get consent for consent in that
way, a huge debate, and it was touch and go actually whether that
whole Biobank programme crashed and burned or whether it could
go ahead at all. Eventually, I believe that the Director-General
of Research and Development was the signatory of those letters
to individuals and the result of that exercise was that of every
2,500 people who were approached, one said "I don't want
my personal records used in this way" but 200,499 basically
said "Fine, how could I possibly object to this." That
is where we are at with the interpretation. The principles are
pretty solid, but I think most of the community involved in making
these decisions is in a defensive mindset. They do not realise
what I see in the clinic every day, which is when I ask patients
questions like this and say "Look, I cannot ask you to be
in a trial because in a way that smacks of coercion, you had better
go away and think." They say, "Doc, what is the matter
with you, you have been looking after me for a fortnight."
Q432 Baroness O'Neill of Bengarve:
In some ways this question has already been answered, which is
the question about the clinical utility of a lot of these exciting
new scientific discoveries. We have been hearing from witnesses
already that there is not very much evidence yet that the new
genome profiling tests have useful healthcare applications. We
have also heard evidence of data in recent publications—for
example on screening for breast cancer susceptibility genes—suggesting
that the new tests might be used for stratifying the population
for breast cancer screening. Where do you think we are on the
clinical utility of these advances?
Professor Sir John Bell: I think you are seeing
just the front edge of what will be quite a significant wave of
activity to try to demonstrate clinical utility. I think the sceptics—who
I have to say were sceptics 15 years ago when I wrote the article,
they are the same guys—will believe that genetics is really
only applicable in diseases where the penetration is 100 per cent.
It is clearly applicable there and nobody disputes it, it is fine,
you can have a little cottage industry that does that stuff—you
know, it is okay, it is not going to have huge effects on the
health service but it will help some patients and that is a good
thing. The real question is whether these tools—and I am
defining genetics very broadly here—will have an application
across a much broader swathe of what we are doing in clinical
practice. The suggestion that the data that comes out of the whole
genome association data, with relatively small but robust odds
ratios, can be used to stratify patients in breast cancer screening
is an interesting idea but we will need to see the data, so I
am the first to say show me the clinical utility, it will be important.
There is already data which is being generated—and we will
see the data within a few months—of transcript profiling
to separate women with breast cancer into high and low risks groups
in a way that you cannot do with other technologies. It may allow
some women who would have been exposed to chemotherapy to be able
to avoid chemotherapy, and other women with bad prognosis disease
who would not have been treated aggressively to be treated aggressively.
Those might be—and I say might because the clinical utility
data is not fully complete—in light of the early data a
very strong marker and we do know that the starting point is that
the way we practise medicine is incredibly inefficient. Only 30
per cent of the people we expose to a new medication respond to
the medication, so how does that work; 70 per cent of people are
not getting any benefit and yet we cannot identify those 70 per
cent when we screen. The whole process of breast cancer screening
is unbelievably inefficient—that is not to say we do not
support it, it is a good thing to do, but it is really inefficient.
Another very good example, which is much more tangible now is
cervical cancer screening. Cervical cancer screening has gone
for years with cytologists looking at pap smears down microscopes,
with about a 50 per cent sensitivity. Jack Cusack has done some
nice work to show that actually the sensitivity is about 50 per
cent so you identify the problem about 50 per cent of the time.
By using genetic tools to look for papilloma virus and the two
strains which are oncogenic, which are 16 and 18, you can rapidly
get to a position where (a) you might be able to eliminate the
pap smear altogether, which would be a significant benefit, but
you also get up to a sensitivity which is nearer 90 per cent,
and it is easier and it is more efficient and you do not have
all the errors you get in cytology labs and you are not running
around chasing what is actually a rather inefficient procedure.
That is here today, in fact the labs will be doing it today, but
it has not been rolled out on a national screening programme anywhere,
but it seems to me inevitable that it is going to happen. I go
from caution, therefore—I think justified caution—about
predicting how much early prediction you can do to the absolute
confidence that some of this will have a big impact on the way
we practise a whole variety of bits of medicine. Let me just remind
you also—because I think this is a really crucial thing—that
the pharmaceutical industry has never really been very interested
in stratifying patient populations, because if you can sell a
drug to a big population, why would you want to sell it to a small
population—in very simple terms that is correct. But the
pharmaceutical industry is also realising that it is actually
quite difficult in the modern world to discover new innovative
medicines that have a very high efficacy signal, which in other
words actually have a beneficial impact on 50, 60, 70 per cent
of the people who receive it. They can only get from where they
are now to where they need to be by stratifying the patient populations.
They are not going to be able to stratify it by asking who eats
cornflakes in the morning, it is going to have to be something
robust and it is going to have to be genetic tools, there is nothing
else that will get them there. We already have a hint in Alex's
game where a drug which was developed by AstraZeneca called Iressa.
It showed promising results in phase two, failed in phase three
because there was a subset of the population that had a genetic
variation in the target, the EGF receptor, who were highly responsive
to the drug, but the people who did not have that genetic variation
were non-responsive. That information is make or break for a new
agent, so my view is that there will be enormous pressure to use
this in those kinds of settings and they may not be the early
prediction of common disease which is, I accept, a goal eventually
of this technology but not proved. It will be in a whole host
of ways that we use these in the clinic in common everyday diseases
and it will, I think, encroach on almost every aspect of medicine.
Q433 Baroness Finlay of Llandaff:
I wonder if I might just follow-up on the comments about pharmacogenetic
data initially and ask you how soon you think some of these tests
are actually going to become available in clinical practice and
be rolled out. I am asking you to crystal ball gaze again.
Professor Sir John Bell: There are two types
of tests, one type is prevention of adverse effects which largely
tests around metabolic function to try and make sure people are
on the right dose and also by trying to identify the idiopathic
adverse effects that you get with some drugs. We are starting
to fill that story in rather more effectively. We know quite a
bit about drug metabolism now. We know about variations in the
P450 cascade. There are available tests in that arena already
that I have to say are not widely used, again for the clinical
utility problem. Nobody can jump that hurdle, and in terms of
idiopathic adverse effects the group in Oxford led by Rory Collins,
who has done probably more work on statins than anyone, has done
a beautiful piece of work which I believe is now in press. The
big adverse effect from statins is a myopathy that can sometimes
be fatal, and given that in this country it looks like we are
going to put everybody on statins over the age of 50 it would
be kind of nice to know if you are going to drop over. It turns
out that there is one gene that is responsible for that. They
know what it is and they know what the variant is; it is a transporter
gene that actually pumps the statins out of the muscle cells and
when it has got a mutation in it, you do not pump the statins
out and you get a myopathy and then you die. Okay, it is rare
but were it cheap would we want to know that information before
we prescribed the statin? I think a clinician would want to know
that, so a lot of this has got to do with how do you get the clinical
utility. Then how do you get it out there and used in an efficient
way, because the last thing a GP wants to do is to decide somebody
needs to go on a statin and then wait a month to get a result
of a test that he cannot interpret anyway. It has to be clean
and tidy and easy for practitioners to actually access information
that gives them good decision-making tools on a clinical pathway.
It seems to me that that piece will come together but there is
a series of hurdles around the clinical utility and implementation
of the information once you actually have it.
Q434 Baroness Finlay of Llandaff:
What about cost? You have not alluded to cost at all.
Professor Sir John Bell: The cost-effectiveness
of screening everybody for a statin variant when such a small
percentage of people have problems will be pretty suspect. I have
never done the analysis, but you can imagine what it would look
like. Rory I guess had 150 patients—he looked at 40,000
in trials so that is roughly what you see. In my view the best
way to do this will be to bundle it; the great thing about genetics
is you can get the answers to all the questions in one test because
you answer all the questions on one chip or you answer all the
questions on one genome sequence, and then you extract the information
you need out of that. If you say there are about 100 things that
would be interesting to know, that would be useful in clinical
practice, and for 1,000 bucks we could sequence the genome and
stick it on a chip and extract the information, then the cost-effectiveness
starts to change quite dramatically because basically you are
bundling all the information a person might need in their entire
lifetime in a single test.
Q435 Baroness Finlay of Llandaff:
If we look at the possibility of getting genomic data versus the
possibility of having gene expression data, then of those types
of tests which do you think is the most mature and which do you
think is likely to come into clinical practice soon?
Professor Sir John Bell: The one that is most
mature is the gene expression data platforms and that is because
they launched early in this guy's business in cancer. There is
now a very substantial amount of data and there are a lot of clinical
utility studies that are almost ready to report—in fact
FDA has actually approved a test for breast cancer based on clinical
utility data, so that is mature. I do not think you need to assume
that because it is mature that will be the definitive way to approach
this because it may well be that you can infer almost all that
genetic expression data from what is present in the germ line.
We know that there is an expression, QTLs, which relate expression
levels to genetic variations in DNA—germ line genetic variations
are well known in all diseases. In fact, it may well be that the
definitive data will come from the DNA and not from the RNA but
in the interim it is clear that RNA data is providing people with
some rather interesting clinical signals.
Q436 Baroness Finlay of Llandaff:
Do you think that DNA data though is always going to need to be
backed up by gene expression data?
Professor Sir John Bell: No. Do not forget there
is both the germ line data and in cancers there is the somatic
data. One of the reasons you get these strange expression profiles
in cancer is because you have all these somatic mutations and
duplications of lesions and bits and pieces and new variants,
so being able to generate that from the tumour itself is actually
quite powerful.
Q437 Baroness Finlay of Llandaff:
Do you think that that type of information will be coming out
of depositaries such as tumour banks or do you think it will all
be coming from patient direct data?
Professor Sir John Bell: Alex, I will pass that
one over to you.
Professor Sir Alex Markham: Some of it will
be coming from tumour banks most definitely, and we already have
cancer genome projects in the UK and your Lordships have been
to see the US equivalent, so there is huge ambition for that.
I have to say that so far their output has been a bit disappointing.
You are not finding consistent changes across similar groups of
tumours when you start looking for consistent themes, but it will
come, so we need tumour banks—and we have those now in the
UK—and a lot of work will be done with individual patient
samples still, which of course is all that tumour banks are. I
think there will be interesting things in pharmacogenetics; some
of the stuff that is out there and in process right now are things
like testing for malignant hypothermia, people who respond badly
to general anaesthetics. We have a big programme looking at polymorphisms
for people's metabolism of Azathrioprine which we still use quite
heavily in rheumatoid arthritis. There is an awful lot of pharmacogenetics
going on in the management of epilepsy because there are paradoxical
responses to some of the most effective new drugs there, and of
course quite a bit of pharmacogenomics in people with HIV, some
of whom have paradoxical side effects to some of the better agents
there. There is a lot coming, therefore, and in terms of how do
you look at value for money and affordability, I think that the
NICE process can be applied to this, there is no reason why not.
It is much maligned but I think NICE does a very, very valuable
job for us in reminding us that there are always other things
to spend your money on.
Q438 Lord Taverne:
In the light of what Sir John says about the value of a single
comprehensive test, if it could be done at a reasonable cost would
you see ultimately that every newborn baby would be subjected
to tests, as I understand is being envisaged in the United States?
Professor Sir John Bell: I think there is an
ethical discussion you have got to have before you start testing
newborn babies. The idea is that at some stage during life—the
genetic data, until you get a cancer where you have got somatic
variations where you have to do it again, the underlying germ
line genetic data is basically stable so you can do it at any
stage in life. That will give you a set of information that might
be varyingly useful at different stages—if you went on this
drug you would look at that variant, if you go on that drug you
need that variant, you have got a disease but what sub-type of
disease and you would look at the variants that might drive that.
My suspicion is that that would be a very efficient way to do
it and one can now see for the first time how you might do that
in the relatively near future.
Q439 Baroness Finlay of Llandaff:
Should we be requiring pharmacogenetic tests in order to optimise
efficacy and reduce side effects in terms of drug development,
should that be a criteria for licensing?
Professor Sir John Bell: My view is that you
are using the market rather effectively to actually make sure
that that happens, because the truth is the NICE process says
if you do not give us a decent signal in efficacy then we are
not going to buy the drug, and one of the only ways to get that
signal up would be to start to look at sub-populations. Can I
just add something which I think is really important? That is
that the one thing we do not do, and we have not done, is that
we have not incentivised pharmaceutical companies post-registration
to find the subset of people in which they are getting a big signal,
so provided they got their ticket, and a lot of drugs of course
just get their ticket, so they are there and they are selling
it across the piece based on a number of qualities, what you would
really like to do is to get those guys to keep working, to say
actually guess what, we have found the 30 per cent of patients
with pancreatic cancer in whom you get all the beneficial results.
If they did that I would put to you it would not be unreasonable
to allow the price of the drug to go up for those patients to
absorb the full quality benefit that is consolidated in the 30
per cent of the population. We do not do that so I think there
are some interesting questions about how you can further incentivise
industry to do this more aggressively in the post-registration
model. Pre-registration I think you have got them actually because
the NICE thing does it for you.
|
